Comprehensive Review on Automated Fruit Disease Detection at Early Stage

Abstract

Fruits are now cultivated in many different countries, which has increased global fruit output to 2,914.27 thousand tons. Numerous countries want to increase their fruit production in the next years, thus the number of countries producing fruits is expected to keep growing. But despite this, a variety of challenges and problems are still experienced while growing crops. These include problems with the fruit's general quality, the cost of manufacturing, the state of the seed, and the fruit's own illness. The main causes of fruit diseases' detrimental impacts are microbes and fungus. Early fruit disease detection is used to foresee fruit disease, which helps farmers save money by lowering the amount of capital they have to spend. To stop fruit illnesses in their early stages, it is crucial to figure out the best way to identify fruit infections. Many studies on a variety of fruits, including the papaya, apple, mango, olive, kiwifruit, orange, etc., have employed deep learning approaches. This study compares several ways for image capture, pre-processing, and segmentation as well as deep learning techniques. The study discovered that the best deep learning strategy for a particular collection of data may change depending on the system's computational power and the data being used. The results of this study show that a convolution neural network is more accurate and can predict a wide range of fruit diseases.